You are on page 1of 10

Energy 54 (2013) 184e193

Contents lists available at SciVerse ScienceDirect

journal homepage:

A novel vertical axis water turbine for power generation from water
J. Chen a, *, H.X. Yang a, C.P. Liu b, C.H. Lau b, M. Lo a
The Hong Kong Polytechnic University, Kowloon, Hong Kong, China
Water Supplies Department, The Government of the Hong Kong SAR, China

a r t i c l e i n f o a b s t r a c t

Article history: In order to help water supply utilities improve water supply management, offer early warning and
Received 10 September 2012 control water pipeline leakage, so as to provide clean water to residents, many instruments are installed
Received in revised form to monitor hydraulic and water quality conditions along the water pipelines. However, one of the
27 December 2012
challenges of such tremendous monitor system is its limited power resources to make the smart sensors
Accepted 31 January 2013
Available online 11 March 2013
and sensor networks operate continuously and safely due to poor site conditions. In this paper, a novel
vertical-axis water turbine has been developed for hydropower harness inside water pipelines when
extra water head can be consumed. The small hydro power generated can be used for power supply to
Water pipeline monitor
the data collection systems in underground and congested locations or to maintenance work at remote
Hydro power locations. The device was developed through CFD (Computational Fluid Dynamic) simulation and lab
Vertical axis water turbine tests. The performance of the developed water turbines and the flow characteristics inside water pipes
Drag-type were simulated using the ANSYS CFD packages: FLUENT and Gambit. After the simulation, a number of
Lift-type prototypes were fabricated and tested to measure the real power output and verify the simulated results.
CFD simulation and experimental test The results indicate that the simulation outcomes can offer a good guidance for the rotor design even
though the difference between the simulation and experimental results is fairly large. The results also
show that the rotor with a hollow structure combined with an eye-shaped slanted block in a pipeline
could generate the maximum power as compared with other types of vertical-axis rotors. Furthermore,
the simulation and test results of different generation turbines offer valuable information for developing
hydropower harness device in confined conditions.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction In terms of water leaking control, Rajani [3] and Kleiner [4]
summarized the physically and statistically based models for struc-
Water is essential for life and plays a vital role in the proper tural deterioration prediction of water mains through model
functioning of the Earth’s ecosystems [1]. Most of our daily water description, critique and data requirements. However, all these pre-
supply is offered by the water utilities which have to ensure the safe diction models need a lot of data with different types. For example,
and clean water supply for local residents. Modern water utilities, the Frost load model [5] requires continuous freezing index, soil
however, are facing several challenges because many urban water properties, thermal gradient at the freezing front, frost depth and
pipelines installed during the rapid urban growth period after the other related variables. Based on the co-author’s [6e8] publications
World War II are now suffering from rapid deterioration [2]. The of the Hadzilacos’s [9] models, factors such as traffic load, frost load,
major challenges involve how to early warn and control water earth load, pipe working and surge pressure, temperature change,
leakage, how to efficiently control and detect water quality, how to internal and external pipe diameter and corrosion coefficients,
manage the pressure distribution along the pipeline, how to bursting strength, tensile strength, maximum external load and
maintain and renew the pipeline and how to optimum design and fracture toughness have been taken into consideration in their
operation of water distribution network. A large number of tech- probabilistic framework. Therefore, Rajani [3] and Kleiner [4] claimed
nologies and prediction models have been developed to deal with that even the best physical analysis model is good only under the
the challenges mentioned above during the past a few decades. circumstance that relevant data are available for its implementation.
Water quality is another rising issue. The United states
* Corresponding author.
Environmental Protection Agency (EPA) claimed that microbes,
E-mail address: (J. Chen). chemicals, and pathogens are still a concern even EPA has set

0360-5442/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
J. Chen et al. / Energy 54 (2013) 184e193 185

standards for about 90 contaminates. Levin [10] suggests that we There are many kinds of conventional water turbines which
need better, more efficient and sensitive monitoring tools and work at different water head and flow rate, such as Francis, Kaplan,
strategies, especially to assess microbial risks and groundwater Bulb and Pelton turbine. And the efficiency of those water turbines
contamination. is normally range from 75 to 95 percent based on the size of hydro
Most of these technologies and models are based on many turbine generator aggregate.
types of data, enough detection nodes and synchronous and According to Fig. 1 [15], Kaplan turbine looks like a promising
continuous data gained at one node for the analysis, inspection turbine for the water pipeline because it can work at low flow rate
and prediction of the water leakage and quality so that water and low water head condition. But Kaplan turbine will alter the
utilities need to install a lot of different types and quantities of pathway of water flow and cannot be well integrated with pipeline.
sensors along pipelines to gain the different data for leaking and Ampair’s UW100 picohydro turbine which is a kind of bulb turbine
contamination inspections at different locations. Till now, most of was proposed by the Saftner [16] to generate the power for the
the sensors are powered by batteries. However, the fact mentioned wireless sensors. Although this kind of water turbine will not
by the Younis [11] is that it is envisioned that monitoring network change the pathway of water flow, the generator and some elec-
always need hundreds of nodes that operate on small batteries. trical components were immersed in the drinking water. Any
According to Yang [12], the operation period per battery charge is improper waterproofing of this system will contaminate the
only 21 day when an aqueous sensor network was applied to drinking water. And the loss of water head of this system also needs
measure environmental parameters of interest for the lake or to be well evaluated. Thus Water Supplies Department (WSD) of the
drinking water reservoir. Power supply to the data acquisition Hong Kong SAR Government has commissioned the Hong Kong
systems, especially the magneto elastic transmitter and sensor Polytechnic University to develop a hydro turbine for pipeline ap-
interface board, is usually off. Hence, they proposed to obtain plications which must satisfy the requirements mentioned above.
power supply from the environment, such as thermoelectric and A vertical axis water turbine system shown in Fig. 2 was pro-
mechanical power. A real wireless sensor network was used by the posed to satisfy above requirements. By using such turbine, the
Stoianov [13] to monitor the Boston city’s water supply. Many pathway of water flow will not change, the electrical parts can be
sensors relied on battery operation due to the fact that many of the arranged outside of the water pipe, and there are some of ready-
sensor locations do not have access to local utility grid power. Even made tubular T joints which can be used directly on the pipeline.
using lossless data compression algorithms, the battery life (6 V After searching the existing vertical axis turbines, lots of researches
12 Ah battery) is about 50e62 days. They claimed that the major focused on the application of the vertical axis turbines in the open
challenge in developing the wireless monitoring system [14] is air or water condition [17e19]. A few of researchers have turned
how to balance the conflict between long distance, communica- their interest to develop the vertical axis turbine used in the half
tion, bandwidth, local data processing and the constraints for low- [20e22] confined condition due to the fact that the Betz limitation
power consumption. can be broken. Several journal papers have been published about
Therefore, more and more studies have focused on how to the vertical axis water turbine working in a complete confined
provide power to the monitoring sensor networks which are the condition. One company named Lucid Energy [23] developed a lift-
essential for the modern water mains management technologies type vertical axis water turbine for the big water pipeline recently.
and methods, especially for the sensors which have no access to Fig. 3 shows this kind of lift-type vertical axis water turbine. And
grid power and need to offer continuous monitoring. In this paper, a this system is only developed for the pipelines which is bigger than
new type of power supply system is proposed, which harvests a 24 inch. Therefore, the fact is that there is very little information
little hydropower from the pipeline for power supply to the data about the vertical axis water turbine used in a confined condition,
acquisition systems. This paper reports the development details of especially for the small pipelines. In this paper, we want to develop
the micro hydropower system including CFD (Computational Fluid the vertical axis water turbine for 100 mm pipeline to convert as
Dynamic) simulation and experiments. much as possible energy when the average water velocity is 1.5 m/s
and the water head drop is less than 5 m.

2. Hydropower harvest inside the pipeline

In order to offer sufficient and safe power for the monitoring

system, abundant hydropower inside pipeline seems very attrac-
tive. The superfluous water head or small part of kinetic hydro- 10
power inside pipeline can satisfy the power requirement of the Pelton Turbines M
10 W
monitoring system. But the challenge of using hydropower is to 0M
select a water turbine which can suit for the requirements of the W
water pipeline condition, especially for the underground and urban 100 10
cities condition. The requirements are listed below: M

1M W
➢ The average water velocity in the pipe is 1.5 m/s. Francis Turbines
➢ Although some superfluous water head may exist in the Crossflow
water pipeline, the water turbine cannot consume a lot of 10
water head to make sure drinking water can overcome pipe 0.1
friction loss and can be delivered safely to the end user.
W Kaplan Turbines
➢ At most of time, space for system installation is very limited in
the underground condition and urban cities. Therefore,
pathway of the water flow cannot be altered. 1 10 Flow (m3/s) 100 1000
➢ The system does not have any underlying dangers that may
affect water quality Fig. 1. Turbine application chart.
186 J. Chen et al. / Energy 54 (2013) 184e193

Fig. 2. The proposed system for the water pipeline.

3. Design of the hydropower system

Fig. 4. Design methodology.

3.1. Design process

design of the vertical axis wind turbine [24e26]. In this paper, the
Due to the development of the CFD, the modern fluid mechan- ANSYS CFD package, which is considered to be one of the best fluid
ical design is changed from pure experimental design to design software packages nowadays, is applied to simulat the ver-
simulation-based design supplemented by experimental test. This tical axis water turbine. To simulate the unsteady flow field, the
design process is described in Fig. 4, which can be divided into two sliding mesh method was adopted. The computational domain was
stages. In the first stage, the proposed water turbine was made divided into internal and external parts. The face between the in-
using the CAD software, then the CAD model was imported into the ternal and external domain is defined as an interface which allows
CFD software to simulate different water turbines’ performance the flow properties to transport through the interface. A simple
under different working conditions. After that, the simulation re- algorithm was used to solve the NaviereStokes equations. The
sults were compared to the design aims to check whether the enhanced wall treatment was chosen for the near-wall region to
performance of the proposed model can meet the design aims. If resolve the viscous layer and the Yþ of the turbine blades should be
the performance of the proposed model cannot meet the design close to 1. The whole computational domain was decomposed by
aims, then the first two processes were repeated till the perfor- 1.5 million unstructured tetrahedral meshes.
mance of the proposed turbine achieves the design aims. The mesh of the final design is presented in Fig. 5. Although the
Once the simulated performance of the proposed model was mesh number remains a controversial issue, the degree of the mesh
close to the design aims, the testing stage was launched. If the number was considered well enough for such a small computa-
performance of the proposed model cannot meet the design aims, tional domain to make an investigation of the grid-dependence
the first and second stage will be repeated until the model meets unnecessary. And the computational time of such one case is
the requirements. There are mainly two purposes of the experi- about one day using a desktop computer with Intel Core2 Quad CPU
ments, i.e. to ensure that the proposed models meet the design Q9550 @ 2.83 GHZ 2.83 GHZ processor. It means that we need four
requirements, and to find whether the simulation is really days to get one turbine’s power coefficient curve at four different
instructive and the difference between the experimental results rotational speeds. That is why some simulations of the water tur-
and simulation results. bines only have one or two power coefficient points.

3.2. CFD settings

The design and prediction capacity of the CFD has been

confirmed to be effective and instructive in many areas, such as
fluid machinery, aerospace and atmospheric environment and
other fields. The CFD method also has been widely used in the

Fig. 3. Lucid energy lift-type turbines. Fig. 5. Detailed mesh of the simulated models.
J. Chen et al. / Energy 54 (2013) 184e193 187

Fig. 6. Whole testing rig.

It is unfortunate that none of turbulence model can be applied to monitoring point of the inlet pressure and outlet pressure were
all the problems. The choice of the turbulence model depends placed upstream 4 times and downstream 8 times diameter of the
mainly on the considerations such as the physics encompassed in pipeline, respectively.
the flow, the required level of accuracy, the available computational
resources, and the amount of time available for the simulation. 3.3. Prototypes and experiment setup
Taking into account the available computing resources, the number
of the design models and relative shorter design cycles, a two- The simulated water turbine was fabricated by a Computer
equation model is chosen to approach the Reynolds-averaged Numerical Control (CNC) machine using stainless steel which will
Navier-stokes equations. The most commonly used turbulence not contaminate drinking water and has enough strength. The
model is the kε turbulence model, and the realizable kε model is prototype water turbine was installed into a lab testing system as
likely to be superior to the standard kε and RNG kε models for shown in Fig. 6 to measure the real performance of the developed
flows involving rotations, boundary layers under strong adverse turbine. The testing system is mainly composed of a water tank of
pressure gradients, separation, and recirculation. Another kind of 1.6 m3, a circulation pump, valves and sensors in a circulation loop.
the two-equation turbulence models is the ku models. Among the The pump is assumed to operate at 50% of its full capacity during
ku models, the SST ku model is more reliable and accurate than the 8e10 h final acceptance test. A 37 kW pump produces 82 m
the standard ku model due to the additional features added into water head at 81 m3/h (equivalent to 2.8 m/s in a 100 mm pipe)
the SST ku model [27]. According to the above analysis, the SST to simulate the real water pipeline conditions. A variable speed
ku model and realizable kε model were adopted to simulate the drive will control the revolution speed of the pump. Several
initial case. The default values in Fluent are used for all the constant valves were used to regulate the water pressure and flow rate.
values of the SST ku and realizable kε model. Velocity inlet was There are also several sensors used to monitor the flow rate,
used for the inlet of pipeline. The velocity of the flow is 1.5 m/s temperatures and pressures inside the pipe. The flow meter
according to the datum provided by the Water Supply Department was placed far away from the water turbine. The inlet and
of the Hong Kong SAR Government. The outlet of pipelines is the outlet pressure sensors were installed upstream 5 times and
pressure outlet boundaries of the model. It is found that there is no downstream 4 times diameter of the pipeline, respectively, due to
significant difference between the simulation results of the two the limited space in our lab. If 5 m water head is consumed
models over a large number of time steps. Therefore, we selected totally, the estimated power is a  384.8 W, where a is the block
the realizable kε model to simulate all the subsequent design ratio which is defined as the ratio between the turbine’s diameter
models. and the diameter of the pipeline. If the water turbine is well
In order to calculate the power output, the torque coefficient of designed, it will produce at least 80w power when the
the water turbine is monitored. The average torque coefficient will assumptive system efficient is 0.3 and a is 0.7. The following
be calculated until the toque coefficient shows periodic variation. equation is used to estimate the potential power generation of
To evaluate the pressure drop caused by the water turbine, the the water turbine:

Fig. 7. Initial proposed designs.

188 J. Chen et al. / Energy 54 (2013) 184e193

Table 2
Simulation and testing results of drag type models.

Case Blade Rotor Angular Power Pressure

No. No. size speed output (W) drop (pa)
(mm) (rad/s)/TSR
Simulated S5 3 66 37.5/0.825 0.543 1551
drag turbine S6 3 86 22.5/0.645 3.8 6777
(solid) 30/0.86 4.28 7030
37.5/1.075 4.2 7268
S7 3 95 22.5/0.7125 10.64 *
30/0.95 12.34 *
37.5/1.1875 13.59 *
45/1.425 14.49 *
Fig. 8. The solid and hollow drag-type turbines. Test result T3-S6 3 86 * 0 0

p1 v21 p2 v22
P ¼ rg z 1 þ þ  z2   Q drag-type turbines. Fig. 8 shows the difference between the solid
rg 2g rg 2g and hollow turbines. All the results of simulated and testing cases
Therefore, a micro permanent magnet power generator (FF- were numbered. A letter S will be added before the simulated case
50 W) was used due to the propriety of low starting torque and zero number. Similarity, a letter T means testing. If one case was simu-
cogging design whose rated power output is 50 W and maximum lated and tested, both S and T will be used in the case number.
output can reach 100 W. In addition, the generator winding would Table 1 shows the simulation results of the four different lift-
be customized finally after testing the system performance, such type vertical water turbines at different rotational speeds. In or-
that the output voltage at the maximum power point can match the der to improve the self-start ability, the lift-type turbine with high
battery charging voltage, which is around 28 V. solidity was used. The rotational speeds of the lift-type turbines
were initially decided according to the principle of the high solidity
vertical axis wind turbine in open conditions. The higher solidity,
4. Experimental and simulation results of the three the sharper the power curve is [30]. In order words, the maximum
evolutions power point of the turbine is close to lower tip speed ratio (TSR).
Thus, the evaluated TSR for the turbine ranged from 0 to 4. The
4.1. First generation of the hydropower system interval between the two TSR points was 0.4. After the simulation,
we found that the power coefficient is negative when the TSR is
According to literature survey, there are very few reports about larger than 2.5. Therefore, the TSR of most of the simulation cases is
the vertical axis water turbine working in a confined condition less than 2.5. The TSR is defined as:
[20,28]. Thus, there is little reference which can be found for this
project. Therefore, in the initial design stage, different spherical u*r
drag and lift-type water turbines which are mainly based on the l ¼
vertical axis wind and water turbines were proposed and simulated
using the CFD software. The airfoil used was the NACA 4415 airfoil Where v is the water velocity, u is angular speed and r is the radius
which is considered as a promising airfoil for the Darrieus rotor of the turbine. The solidity is expressed as:
[29]. The drag-type turbine was a conventional Savonius rotor with
three blades and zero overlap. Some of the proposed models (left) N*c

and real models (right) are shown in Fig. 7. D
The design parameters of the lift-type turbines are their chord
length and blade number. The rotor size and blade number are the Where N is the blade number of the turbine, c is chord length of the
design parameters of the solid drag-type turbines. The diameter of airfoil and the D is the diameter of the rotor.
the hollow circular cylinder is the design parameter of the hollow Comparison between the case S1 and case S2 shows that the
chord length influence on the turbine performance is obvious. The
conclusion drawn from the comparison is the same as that of a
Table 1 vertical axis wind turbine in the open condition. The higher solidity,
Simulation and testing results of the lift-type model. the sharper the power curve is. The reason why the working
Case Blade Chord Angular Power Pressure principle is the same between a wind turbine and a water turbine
No. No. length/ speed/TSR output drop (pa) may be that the rotor blockage value is not big enough to influence
solidity (rad/s) (W)
Simulated S1 3 24/1.09 18.75/0.4 0.65 1101 Table 3
lift turbine 37.5/0.8 1.86 682 Simulation and testing results of drag type models.
diameter 56.25/1.2 3.14 1001
Case Blade Rotor Angular Power Pressure
(66 mm) 75/1.6 1.28 915
No. No. size speed output (W) drop (pa)
S2 3 32/1.45 18.75/0.4 0.48 749
(mm) (rad/s)/TSR
37.5/0.8 1.568 927
56.25/1.2 0.22 * Simulated S8 3 66 37.5/0.825 0.543 1551
75/1.6 0.027 2495 drag turbine S9 5 66 18.75/0.4125 1.875 7829
S3 5 16/1.21 37.5/0.8 1.4 3025 (solid) 28.125/0.618 3.9375 8173
56.25/1.2 1.65 3342 37.5/0.825 0.85 *
S4 6 24/2.18 37.5/0.8 1.14 2223 S10 10 66 150.33 3.67 7752
56.25/1.2 2.06 2560 22.5/0.495 4.63 7344
Testing T1-S1 3 24/1.09 * 0 * 30/0.66 5.24 7086
result T2-S2 3 32/1.45 * 0 * 37.5/0.825 4.8 6833
J. Chen et al. / Energy 54 (2013) 184e193 189

Table 4 The reason for this increase may be that the numbers of advancing
Simulation and testing results of drag type models. blades which generate the thrust forces is increased in the confined
Case Blade Diameter Rotor Angular Power Pressure condition. And the drag force produced by the returning blades is
No. No. of hollow size speed output drop (pa) reduced due to the smooth pass of the incoming flows. Three tur-
circular (mm) (rad/s)/TSR (W)
bines with different diameters of hollow circular cylinder were
presented in Table 4. It is found that the power output is almost
close to zero when the water velocity is 1.5 m/s and the pressure
Simulated S11 10 20 92 15/0.46 3.06E-04 137,200
drag 30/0.92 9.65E-05 137,200 drop is very high. One possible reason, which is responsible for this
turbine S12 10 30 92 15/0.46 2.92E-04 95,746 phenomenon of lower power output and high pressure consump-
(hollow) 30/0.92 1.01E-04 97,314 tion, is that the angular speed of the rotor is evaluated improperly.
40/1.22 3.31E-05 * The optimum rotational speed of this hollow rotor with multiple
50/1.53 1.69E-04 *
S13 10 40 92 15/0.46 1.01E-04 68,600
blades may be much lower than the rotational speed of estimation.
30/0.92 2.59E-04 71,736 Hence, this rotor is working at off-design condition which causes
turbulence flow in and around turbines. Another reason may be
that this type of turbine is not suit for this design conditions.
the working principle of the vertical axis turbine in the open con- Comparison between Tables 1e4 shows that the power output
dition. The comparison of the case S3 and case S4 shows the blade of the drag-type turbine is generally higher than that of the lift-type
number influence on the turbine performance. The power output of turbines and the same for the pressure drop in a confined condition
the case S3 and case S4 is much lower than that of the case S1. except for the hollow drag-type turbines, which indicates that the
However, the pressure drop of the case S3 and case S4 is much solid drag type especially ones with more blades and larger
sharper than the case S1 and case S2, although the pressure drop of diameter, is more suitable than the lift-type in a confined space.
the case S2 is not shown at angular speed 56.25 rad/s because the Although the power output of the first generation of the hydro-
datum was lost one year ago due to the damaged of a hard disk. The power generators is sufficient enough for the power of the moni-
case S1 and case S2 were fabricated and installed into the pipeline toring sensors, there is still a lot of room for improvement due to
to measure their real power outputs. The last two rows in Table 1 the pressure drop of the initial proposed turbine which is much
show the testing results of the case S1 and case S2. Zero power lower than the allowed pressure head consumption of 5 m. Further
output is predictable when the case S1 and case S2 were used to modification is necessary for the drag-type turbines so as to
drive the 50 W power generator. generate more power.
The optimum TSR of the drag-type turbines is about 0.8 in the
open condition. Thus, the angular speeds of the drag-type turbines 4.2. Second generation of the hydropower system
can be chosen when the TSR is close to 0.8. Table 2 shows the in-
fluence of the rotor size on the power output of the solid drag-type Based on the conclusions drawn from the development of the
turbines. It is obvious that the power output increase with the in- first generation turbines, more attention was paid to the drag-type
crease of the rotor size. The highest power output is 14.49 W for the turbines with more blades and larger size on their second gener-
case S7. The power output increase could be probably attributed to ation as more water head need to be consumed to have more power
several factors. The power output is in direct proportion to the output. According to the working principle of the drag-type rotor, it
sweep area of the turbine in the open condition. It seems that this is obvious that the returning blades of the drag type turbine create
principle is applicable to confined condition. However, the power most of the resistance force. If we can install a block inside the
changes non-linear along the sweep area. A small change of the pipeline to surround the returning blades, the water flow will not
sweep area could cause big changes of power output. The second attack the returning blades directly. At the same time, part of the
factor is the block ratio which is only used in the confined condi- water head energy is converted into kinetic energy due to nar-
tion. Water flow can skirt around the turbine in the open condition, rowing the water flow channel. The returning blades, advanced
but it can only act on the water turbine directly due to the limita- blades and some blocks are shown in Fig. 9.
tion of the pipe wall, thus generating more thrust force. Therefore, four types of blocks were designed for solid drag-type
The influence of the blade number on the power output of the turbines. Fig. 10 shows the four types of blocks according to
solid drag-type turbines is presented in Table 3. It is clear that the sequence of the evolution. In order to install the block through the T
more blades a turbine has the more power it can generate. The joints, the size and type of the blocks are restricted. From left to
maximum power output is 5.24 W when ten blades are adopted for right in Fig. 10, the vertical block is the first proposed block which
the solid drag-type turbine working at 30 rad/s. In the open con- can directly be inserted into the pipeline through a T joint. The
dition, the power coefficient will decrease with the increase of the second block which is a long slanted block needs to be divided into
blade number. However, it is opposite in the confined condition. three pieces for installation. Hence, the installation process of the

Fig. 9. The working principle of the drag-type turbine.

190 J. Chen et al. / Energy 54 (2013) 184e193

Fig. 10. Four types of the blocks.

second block is relatively complex. Thus, the third block which is a the principle (more blades for more power) is suitable for the
short slanted block was developed to replace the second block. vertical block when the block ratio is 0.5. The corresponding tested
The last block is a slanted block with an eye-shaped opening. The turbine is the case T5 which has zero power output. More tested
definition of the block ratio which is shown in Tables 5 and 6 is cases with vertical block were listed in Table 6. The increase of the
the same for the vertical, long slanted and short slanted blocks. The block ratio leads to power output increase. The maximum power
block ratio is defined as d/D which is presented in Fig. 9. The block output and pressure drop are 13.4 W and 3.75 m, respectively,
ratio of the eye shaped block is defined as the ratio of outlet area to when the blade number is six and block ratio 0.9. As the power
inlet area. The angular speed of the simulated turbines was eval- output is much less for the vertical block arrangement, a different
uated based on the first generation turbine. block was developed and simulated. Based on the test results of
The simulation results of the vertical and slanted blocks are the vertical block, the simulation of the second block focused on
listed in Table 5 and the test results are presented in Table 6. Case the high block ratio. Although the simulation results shows that the
S14 and case S15 were compared to investigate the blades number maximum power output is 67.69 W when the block ratio is 0.8, the
influence on the turbine performance when the block ratio is 0.5. It test result of power output is only 11.8 W.
is found that power output of case S14 is less than that of the case It is a little complex to install the long slanted block through the
S15 when the vertical block is combined with solid turbine. The T joint therefore a shorter slanted block was developed and tested.
maximum power output of the case S14 is 10.6 W, which seems that It seems that the power output of the short slanted block is close to

Table 5
Simulation results of drag type models.

Case No. Blade No. Block type block ratio Rotor Angular Power Pressure
size (mm) speed (rad/s) output (W) drop (pa)
Simulated results of drag turbine (solid) þ vertical S14 5 Vertical block 0.5 86 37.5 2.2 4007
block or long slanted block 56.25 10.6 5726
S15 10 Vertical block 0.5 86 15 3.67 *
22.5 4.63 *
30 5.24 7305
37.5 4.8 6994
S16 5 Long slanted block 0.6 86 30 12.56 36,829
40 13.77 36,339
50 12.29 36,004
S17 5 Long slanted block 0.8 92 30 55.15 84,523
40 64.68 85,176
50 67.69 80,875
S18 6 Long slanted block 0.8 92 37.5 59.79 87,623
S19 10 Long slanted block 0.8 92 40 64.43 85,329

Table 6
Testing results of drag type models.

Case No. Blade No. Block type Rotor size (mm) Max Power Pressure
output (W) drop (pa)
Testing results drag turbine T4 3 Vertical block 0.5 86 0 *
(solid) þ four kinds of blocks T5-S14 5 Vertical block 0.5 86 0 *
T6 5 Vertical block 0.7 86 0 *
T7 5 Vertical block 0.8 86 1.9 21,952
T8 6 Vertical block 0.7 86 1.1 20,678
T9 6 Vertical block 0.8 86 4.0 23,226
T10 6 Vertical block 0.9 86 13.4 36,750
T11-S17 5 Long slanted block 0.8 92 11.8 *
T12 5 Long slanted block 0.9 92 0 *
T13 5 Short slanted block 0.8 92 12 *
T14 5 Slanted block with eye-shaped opening 0.9 92 24.8 54,880
T15 6 Slanted block with eye-shaped opening 0.8 92 6.1 32,928
T16 6 Slanted block with eye-shaped opening 0.85 92 26.2 46,550
T17 6 Slanted block with eye-shaped opening 0.9 92 32.2 57,036
T16 6 Slanted block with eye-shaped opening 0.95 92 26.6 102,018
J. Chen et al. / Energy 54 (2013) 184e193 191

Fig. 11. Two different combinations.

that of the long slanted block. Three blocks mentioned above have a Table 7 shows the simulation results. The power output of the
common disadvantage, i.e. the offset between the flow pathway case S22 is twice higher than that of the case S19 which produces
and the turbine’s advanced blade blocking will increase when the the highest power in the second generation. The only difference
block ratio increases. Thus, the fourth generation block was between the two cases is the turbine’s hollow structure. The
developed. Through this new block, the flow can act directly on the simulation results indicate that the optimum diameter of the hol-
turbine’s advanced blades. Fig. 10 illustrates this improvement low circular cylinder is close to 45 mm. From the conclusion of the
clearly. The eye shaped block was manufactured and tested without second generation development, the best block is the slanted eye
further simulation. From Table 6, the maximum power of the shaped turbine. Thus, the hollow turbines were used together with
slanted eye shaped block can reach 32.2 W. This is a significant a slanted eye shaped block. Case S23 and case S24 show the in-
improvement and more water head was consumed. The optimum fluence of the turbine diameter with a hollow circular cylinder.
block ratio is 0.9. Only two kinds of diameters were investigated of the slanted eye
shaped block due to time consideration and some indications given
from the simulation of the long slanted block system. The influence
4.3. Third generation of the hydropower system of blade number can be found through the comparison among the
case S24, case S25 and case S26. The simulation results show that
In fact, the power output from the above turbines is enough to the optimum blade number is close to 15. In fact, the difference of
supply the monitoring system. However, there is still much room the simulation power output between the 12-blades turbine and
for improvement according to the estimation. From the above 15-blades turbine is only 2 W. Although the combination of the
simulation and tests, it is found that the solid drag turbine can hollow turbine and slanted eye shaped block provides a significant
produce more power for the first generation of the water turbine improvement of power output, the water head loss is close to 12 m
and the slanted eye shaped block is the best block in the second which is not acceptable. In order to validate the conclusion about
generation. After testing different design combinations of the solid the blade number influence, the turbines with large blade number
turbines and blocks, the only combination which is not tried is the difference were developed. The test results presented in Table 8 are
hollow turbine with a block due to the zero power output of the very inspiring. The maximum power output of this combination
hollow drag type turbines in the first generation. reached 88.2 W and the consumed water head is within the 5 m
The simulation of this new combination is thus carried out. Two limit. The optimum blade number of the tested turbines is 12 which
types of the combinations are presented in Fig. 11. One is the have also validated the conclusion drawn from the simulation re-
combination of a long slanted block and a hollow drag type turbine. sults, although the power output and pressure drop from the
Another one is the combination of the slanted eye shaped block and simulation are much higher than the power output and pressure
hollow drag type turbine. drop of the test results.

Table 7
Simulation results of drag type models.

Case Blade Block type Diameter of hollow Rotor Angular Power Pressure
No. No. circular cylinder (mm) size speed (rad/s) output (W) drop (pa)
Simulated results of drag S20 10 Long slanted block 0.8 20 92 40 99 72,787
turbine (hollow þ long slanted S21 10 Long slanted block 0.8 30 92 40 116.3 74,739
block or slanted eye shaped block S22 10 Long slanted block 0.8 45 92 40 127.5 75,647
S23 12 Slanted block with eye-shaped opening 0.9 45 92 40 226.8 124,337
S24 12 Slanted block with eye-shaped opening 0.9 55 92 40 215.3 124,369
S25 15 Slanted block with eye-shaped opening 0.9 45 92 40 228 122,469
S26 18 Slanted block with eye-shaped opening 0.9 45 92 40 221 122,642

Table 8
Testing results of drag type models.

Case No. Blade Block type Diameter of hollow Rotor Angular Power Pressure
No. circular cylinder (mm) size (mm) speed (rad/s) output (W) drop (pa)
Tested results of drag T17 8 Slanted block with eye-shaped opening 0.9 45 92 66 66.4 43,904
turbine (hollow) þ slanted T18-S23 12 Slanted block with eye-shaped opening 0.9 45 92 74 88.2 47,530
eye shaped block T19 24 Slanted block with eye-shaped opening 0.9 45 92 63 67.4 43,414
192 J. Chen et al. / Energy 54 (2013) 184e193

model. The possible reason for the difference is that the energy
losses caused by water flow friction and various mechanical fric-
tions were not considered in the simulations, and the mismatch
between the turbine’s power curve and generator’s power curve
may exist. Besides, the simulated models and water pipe are not
exactly the same as the tested models and water pipe in shape. In
addition, the effect of the CFD settings on the simulation results is
not well investigated profoundly due to the high cost of compu-
tational effort and the limitation of design time.

5. Conclusions

An inline vertical axis water turbine is developed in this paper

for power supply to water pipeline data monitoring systems. This
turbine system is designed for a 100 mm pipeline. The expected
power output is larger than 80 W when water velocity is 1.5 m/s
and pressure drop is less than 5 m. After simulating and testing
more than 20 different kinds of water turbines under various
Fig. 12. The hydropower system is running on site.
working conditions, the combination of the vertical axis water
turbine with a hollow shaft and a short slanted eye shaped block
4.4. Summary of the three generations was proved to be the best for generating the highest amount of
power output which is 88.2 W. Further improvement can still be
A summary of the typical vertical axis water system is presented made for this novel inline vertical axis water turbine although the
in this section. The power output is increased significantly from final design can meet the expectations under certain limitation. The
zero to 88.2 W during the one-year experiment test and CFD design parameters which need to be further investigated can be the
simulation. The typical designs of the different generations are tilted angle of the turbine blades, folium of turbine blades and the
listed below. hollow shapes of the turbines. Lastly, the simulated and testing
The combination of the hollow drag type turbine and the slan- results offer valuable insight into hydropower harvest inside the
ted eye shaped block not only can produce the estimated power to water pipeline of different sizes, although the system developed in
supply the monitoring system, but also satisfy the power supply this paper focused on the 100 mm pipeline. Some of the turbines of
requirements. The eye shaped block and hollow turbine can be the first and second generations may also be adopted when power
inserted easily through the T joints on the pipeline. The power supply demanding is low.
output and pressure drop can be controlled through the adjustment Several final systems have been installed at designated areas in
of the block ratio, which makes the system readily suitable to the Hong Kong for the purpose of further on-site investigation. One of
different requirements. these systems is presented in Fig. 12. The measured value of the on-
site power output is very close to the tested value in laboratory.

Drag type turbine design Max Head Acknowledgments

power (W) drop (m)
1st gen: 5-blade solid 0 0
The work described in this paper was supported by the Water
turbine (B86 mm) þ
vertical half block Supplies Department and the Inter-Faculty Research Grant of The
Hong Kong Polytechnic University (Project No.: G-YG08). The PolyU
Technology & Consultancy Company Limited provided a lot of
1st gen: 5-blade solid 12.0 N/A laboratory testing support to the project.
turbine (B92 mm) þ 80%
short slanted block

2nd gen: 6-blade solid 32.2 5.82
turbine (B92 mm) þ 90% [1] Learn the issues: water. United States Enviromental Protection Agency; 2012.
slanted eye shaped block [2] Brothers K. A practical approach to water loss reduction. Water 2003;21:
[3] Rajani B, Kleiner Y. Comprehensive review of structural deterioration of water
mains: physically based models. Urban Water 2001;3(3):151e64.
3rd gen: 12-blade (B92 mm) 88.2 4.85
[4] Kleiner Y, Rajani B. Comprehensive review of structural deterioration of water
hollow turbine þ 90% slanted mains: statistical models. Urban Water 2001;3(3):131e50.
eye-shaped block [5] Rajani B, Zhan C. On the estimation of frost loads. Canadian Geotechnical
Journal 1996;33(4):629e41.
[6] Camarinopoulos L, Chatzoulis A, Frontistou-Yannas S, Kallidromitis V. Struc-
tural reliability of water mains. In: Proceedings of the probabilistic safety
assessment and management 96: ESREL 96-PSAM-III; Crete, Greece, 1996(a).
Although the simulated and testing results demonstrate large [7] Camarinopoulos L, Pampoukis G, Preston N. Reliability of a water supply
network. In: Proceedings of the probabilistic safety assessment and man-
quantitatively difference, the trends and variations of the simulated
agement 96: ESREL 96-PSAM-III; Crete, Greece, 1996(a).
results are consistent with that of the testing results from the [8] Camarinopoulos L, Chatzoulis A, Frontistou-Yanas S, Kalidromitis V. Assess-
qualitative point of view. Both the simulation and test results show ment of the time-dependent structural reliability of buried water mains.
that the rotor with hollow shaft and short slanted block can Reliability Engineering and Safety 1999;65(1):41e53.
[9] Hadzilacos T, Kalles D, Preston N, Melbourne P, Camarinopoulos L,
generate much more power than the others. The simulation results Eimermacher M, et al. UtilNets: a water mains rehabilitation decision-support
do offer important theoretical guidance for the design of a real system. Computers, Environment and Urban Systems 2000;24(3):215e32.
J. Chen et al. / Energy 54 (2013) 184e193 193

[10] Levin RB, Epstein PR, Ford TE, Harrington W, Olson E, Reichard EG. US drinking [19] Islam M, Ting DSK, Fartaj A. Aerodynamic models for Darrieus-type straight-
water challenges in the twenty-first century. Environmental Health Per- bladed vertical axis wind turbines. Renewable and Sustainable Energy
spectives 2002;110(Supplement 1):43. Reviews. 2008;12(4):1087e109.
[11] Younis M, Akkaya K. Strategies and techniques for node placement in wireless [20] Kirke BK. Tests on ducted and bare helical and straight blade Darrieus
sensor networks: a survey. Ad Hoc Networks 2008;6(4):621e55. hydrokinetic turbines. Renewable Energy 2011;36(11):3013e22.
[12] Yang X, Ong KG, Dreschel WR, Zeng K, Mungle CS, Grimes CA. Design of a [21] Ponta F, Dutt GS. An improved vertical-axis water-current turbine incorpo-
wireless sensor network for long-term, in-situ monitoring of an aqueous rating a channelling device. Renewable Energy 2000;20(2):223e41.
environment. Sensors 2002;2(11):455e72. [22] Furukawa A, Watanabe S, Matsushita D, Okuma K. Development of ducted
[13] Stoianov I, Nachman L, Whittle A, Madden S, Kling R. Sensor networks for Darrieus turbine for low head hydropower utilization. Current Applied
monitoring water supply and sewer systems: lessons from boston. In: Pro- Physics 2010;10:S128e32.
ceeding of the 8th annual water distribution system analysis symposium, [23] How it works. Portland: Lucid Energy.
Cincinnati, OH 2006, p. 1e17. [24] D’Alessandro V, Montelpare S, Ricci R, Secchiaroli A. Unsteady aerodynamics
[14] Stoianov I, Nachman L, Madden S, Tokmouline T, Csail M. PIPENET: a wireless of a savonius wind rotor: a new computational approach for the simulation of
sensor network for pipeline monitoring. In: ACM/IEEE international confer- energy performance. Energy 2010;35(8):3349e63.
ence on information processing in sensor networks. Massachusetts, USA: [25] Raciti Castelli M, Englaro A, Benini E. The Darrieus wind turbine: proposal for
(MIT) Cambridge; 2007. p. 264e73. a new performance prediction model based on CFD. Energy 2011.
[15] Water turbine chart; 2012. [26] Mohamed MH, Janiga G, Pap E, Thévenin D. Multi-objective optimization of
[16] Saftner DA, Hryciw RD, Green RA, Lynch JP, Michalowski RL. The use of the airfoil shape of wells turbine used for wave energy conversion. Energy
wireless sensors in geotechnical field applications. In: Proceeding of the 15th 2011;36(1):438e46.
annual great lakes geotechnical/geoenvironmental conference, Indianapolis, [27] FLUENT A. 12.0/12.1 documentation users guide manual. Ansys Inc; 2009.
IN; 2008. [28] Antheaume S, Maitre T, Achard JL. Hydraulic Darrieus turbines efficiency for
[17] Kjellin J, Bülow F, Eriksson S, Deglaire P, Leijon M, Bernhoff H. Power coeffi- free fluid flow conditions versus power farms conditions. Renewable Energy
cient measurement on a 12 kW straight bladed vertical axis wind turbine. 2008;33(10):2186e98.
Renewable Energy 2011;36(11):3050e3. [29] Kirke BK. Evaluation of self-starting vertical axis wind turbines for stand-
[18] Kirke BK, Lazauskas L. Limitations of fixed pitch Darrieus hydrokinetic alone application [PhD]. Griffith Universtiy Gold Coast Campus; 1998.
turbines and the challenge of variable pitch. Renewable Energy 2011;36(3): [30] Paraschivoiu I. Wind turbine design: with emphasis on Darrieus concept.
893e7. Presses Inter Polytechnique; 2002.